In STAAD, a surface entity is an object that represents a collection of elements. When the program goes through the analysis phase, it subdivides the surface into a number of elements by itself. This process is called meshing or mesh generation. But during the modeling phase, the user does not see any of those elements which keeps the model small and easy to manage. An entire wall is hence represented by just a few "Surface" entities, instead of hundreds of plate elements.
The advantage of using surfaces is that the minute details involved in the process of converting a physical object like a wall or slab into the mathematical model which consists of plate elements is something that the user doesn't have to bother with. However, in many situations, not knowing these details can lead to errors, some of which are impossible to detect because the underlying elements cannot be seen graphically. Also, not knowing how many elements will be generated during the meshing process can lead to unwanted consequences such as a massive increase in the size of the model, to a point where the program simply cannot handle such a massive volume of data. Consequently, we recommend that surfaces be not used as much as possible.
There is however one situation where the user has no choice but to use surfaces. That is when he/she wants to perform reinforced concrete design of a shearwall per the ACI, British or Indian codes. STAAD can perform a shearwall design if and only if that wall is modeled using
STAAD's SURFACE entity.
When you use surfaces in your model, the resulting mesh of finite elements is not easily
apparent. In fact, the only place in the program where you get a good view of the meshed state of the surface is in the post-processing mode in the node displacement page.
The inability to view the meshed state of the surface makes it very difficult to see the flaws in the modeling. In your structure for example, there may be members which are sharing space with the surface. Unless those members get meshed automatically along with the surface, and more importantly, unless the surface and member meshing is done in such a manner than all the resulting pieces are connected to each other properly, there is no assurance that the members and surface will be able to transfer loads to each other. The inability to see their meshed condition makes it difficult to see if the meshing is done properly. Replacing the surface with quadrilateral elements, and meshing those element is a way to get around the problem. So when you wish to model a floor slab, roof or a wall that does not need to be designed as a shear wall, we
recommend that you use plate elements. In our view, the only drawback of using plates is that you need several of them to model a wall or a slab, and that increases the size of the input data.
But at least you can detect and remove every error in the model, and know exactly how many elements and nodes there are in that model.
Some aspects to consider when analysing mat foundations using STAAD.Pro
Created by Kris Sathia, Technical Support Group, Bentley Systems, Inc.
Question:
I am using STAAD.Pro to analyze a mat foundation that supports a steel frame. I have used the PLATE MAT command and set the springs to COMPRESSION only. I encounter 2 problems :- a) Instability warnings b) Some springs still go into tension
How do I go about correcting these problems?
Answer:
In order to solve problems of this type, there are two aspects that one needs to pay attention to.
1. Solving pure component load cases
Let us assume that there are 4 primary load cases. Load cases 1 and 2 are vertical load cases (downward loads along GY), 3 is wind along X (Nodal forces along global X), 4 is wind along Z (Nodal forces along global Z).
Consider the load cases 3 and 4. They contain only lateral forces. Cases such as these are called component-type primary cases because they represent only one type of load - wind only, seismic only, etc. In the real world, a component type load acts in conjunction with other cases such as gravity. Thus, solving a pure component case is useful only when the results for that case can be combined with those from gravity and other load cases for a steel design for example. But this requires the principle of superposition to be applicable for that analysis run. The principle of superposition states that Results of case A + Results of case B = Results of (case A + case B)
This principle works only for static linear conditions.
Load cases 3 and 4 produce an overturning moment. To satisfy equilibrium, a restoring moment needs to be mobilized to counteract this overturning moment. In these load cases, in addition to the horizontal forces, had there been loads acting vertically downwards, they would have contributed towards a restoring moment. But such loads are not present.
So, the contribution to restoring must come solely from the soil springs. Normally, this is formed when some springs go into compression and others go into tension thus forming a couple. But, by declaring the soil springs as compression-only, the tensile force in the springs are disallowed. Thus, a restoring moment cannot be formed solely by the spring forces. It leads to the instability warnings.
Consequently, a component load case containing only horizontal forces that produce an overturning moment cannot be resisted by a system of soil springs in which a tensile force in the soil spring is disallowed.
There are two ways to address this situation.
1. Instruct the program to "not" solve such load cases. The means to do that is to convert them into
"REFERENCE LOAD DATA" which is explained in section 5.31.6 of the Technical Reference manual.
2. Instruct the program to solve load cases 1 through 4 without the "COMPRESSION" attribute. That attribute can be specified later when the combination cases are solved, as explained next. The PERFORM ANALYSIS and CHANGE commands should also be specified after load case 4. Since the results of these component cases are of no use to us, we will dis-regard them. Thus, we will have something like this.
MEMBER PROPERTIES
1 TO 126 PLATE MAT DIRECTION Y SUBGRADE 12.0 LOAD 1 DEAD LOAD
2. Solving combination load cases
The second aspect to consider in these models is the type of command to use when combining the individual cases to form combination cases. Most people typically use the syntax LOAD COMBINATION nnn. This instructs STAAD to find the result for a combination case by adding the results of the
component primary cases that make up that combination case. In other words, LOAD COMBINATION 7
1 1.0 2 1.0
instructs the program to fetch the results of load cases 1 and 2 and add them algebraically to produce the results of load case 7. An analysis involving the equation [K]{d}={P} is not being done when this syntax is used.
This approach is OK to use only under linear conditions in which the laws of superposition work.
In non-linear conditions such as when COMPRESSION-only springs are present in the model, this principle does not work.
So, instead of adding the results, the program must be instructed to form a physical load case called 7 which contains the load items of cases 1 and 2 factored by 1.0 and perform the operation
[K]{d}={P} where {P} is load case 7. This is the correct way to analyze the model for the loads of cases 1 and 2 acting simultaneously.
The means to achieve it is to use the REPEAT LOAD syntax instead of the LOAD COMBINATION syntax.
For example, the commands
LOAD COMB 7 COMBINATION LOAD CASE 7 1 1.0 2 1.0
LOAD COMB 8 COMBINATION LOAD CASE 8 1 1.0 2 1.0 3 1.0
LOAD COMB 9 COMBINATION LOAD CASE 9 1 1.0 2 1.0 4 1.0
LOAD COMB 12 COMBINATION LOAD CASE 12 1 1.0 4 1.0
LOAD COMB 13 COMBINATION LOAD CASE 13 1 0.67 2 0.67 3 0.67
LOAD COMB 14 COMBINATION LOAD CASE 14 1 0.67 2 0.67 4 0.67
must be changed to
LOAD 7 COMBINATION LOAD CASE 7 REPEAT LOAD
1 1.0 2 1.0
LOAD 8 COMBINATION LOAD CASE 8 REPEAT LOAD
1 1.0 2 1.0 3 1.0
LOAD 9 COMBINATION LOAD CASE 9 REPEAT LOAD
1 1.0 2 1.0 4 1.0
LOAD 12 COMBINATION LOAD CASE 12 REPEAT LOAD
1 1.0 4 1.0
LOAD 13 COMBINATION LOAD CASE 13 REPEAT LOAD
1 0.67 2 0.67 3 0.67
LOAD 14 COMBINATION LOAD CASE 14 REPEAT LOAD
1 0.67 2 0.67 4 0.67
Additionally, since the COMPRESSION-only attribute was removed for cases 1 to 4, the SUPPORTS need to be re-specified just before load case 7 and the COMPRESSION attribute must be included in that command. The data will then look like
LOAD 4 WIND IN Z
1 TO 126 PLATE MAT DIRECTION Y SUBGRADE 12.0 COMPRESSION LOAD 7 COMBINATION LOAD CASE 7
REPEAT LOAD 1 1.0 2 1.0 ..
..
..
LOAD 14 COMBINATION LOAD CASE 14 REPEAT LOAD
1 0.67 2 0.67 4 0.67 PERFORM ANALYSIS CHANGE
LOAD LIST 7 TO 14 ..
PARAMETER
CODE AISC UNIFIED 2010 METHOD ASD
..